17:03:04 <asselin> #startmeeting third-party
17:03:05 <openstack> Meeting started Tue Jan 12 17:03:04 2016 UTC and is due to finish in 60 minutes.  The chair is asselin. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:03:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:03:09 <openstack> The meeting name has been set to 'third_party'
17:03:20 <ja3> moin moin asselin
17:03:31 <asselin> hi, anyone here for 3rd party working group meeting?
17:03:34 <asselin> hi ja3
17:03:45 <ja3> 3rd party, aye
17:03:45 <mmedvede> o/
17:03:51 <asselin> happy new year everyone
17:03:56 <mmedvede> +1
17:04:17 <asselin> #link agenda https://wiki.openstack.org/wiki/Meetings/ThirdParty#Agenda_for_next_Working_Group_meeting
17:04:39 <asselin> #topic announcements
17:04:48 <asselin> anyone have any announcements?
17:05:19 <asselin> I would like to share that the 3rd party ci docs are now officially published
17:05:28 <asselin> #link published 3rd party ci docs: http://docs.openstack.org/infra/openstackci/third_party_ci.html
17:06:33 <asselin> #topic ci-watch
17:07:10 <mmedvede> yay for ci docs!
17:07:52 <mmedvede> I did update puppet-ciwatch,
17:07:56 <asselin> yeah, I think it's a much nicer read :)
17:08:09 <mmedvede> #link initial puppet-ciwatch https://review.openstack.org/#/c/238606/
17:09:11 <ja3> sounds like the review gauntlet being thrown down again there, asselin
17:09:31 <asselin> yeah, looking forward!
17:10:14 <asselin> mmedvede, +1
17:10:43 <asselin> mmedvede, any other open reviews?
17:11:36 <mmedvede> asselin: not yet. Back from a long vaca :)
17:12:13 <mmedvede> but I am working on adding unit testing. it is proving to be hard
17:12:47 <asselin> mmedvede, do you want to share a wip patch? what are the challenges?
17:14:25 <mmedvede> asselin: I have started refactoring config implementation first. Could push wip, definitely
17:14:36 <mmedvede> I mentioned before, as is, the code is not testable
17:15:06 <mmedvede> maybe config testing is too low level, and we should add more useful tests, e.g. to the event parsing
17:15:25 <asselin> yeah, I'd like to help out without duplicating effort
17:19:32 <mmedvede> it is circular - hard to refactor without tests, and hard to write tests for existing code. Could start with failing test cases probably
17:20:49 <asselin> I think that's fine.
17:21:04 <asselin> we could create a feature branch if you prefer to develop it there
17:23:04 <mmedvede> are we able to create it ourselves?
17:23:21 <asselin> probably not, but we can ask in infra
17:26:34 <mmedvede> asselin: also on ciwatch, I do not mind this to merge: https://review.openstack.org/#/c/258125/
17:28:05 <asselin> mmedvede, ok, then I think we should merge it.
17:28:58 <asselin> #action asselin to request a feature branch for ci-watch to refactor for testability
17:29:23 <asselin> mmedvede, anything else?
17:29:30 <mmedvede> no
17:29:49 <asselin> ok I did create some blueprints for ci-watch
17:30:16 <asselin> #link ci-watch blueprints: https://blueprints.launchpad.net/ciwatch
17:30:50 <asselin> i kept these minimal so we can track and develop them further using launchpad instead of the wiki page.
17:31:32 <asselin> but first priority is testability imo
17:31:37 <apoorvad> hi
17:31:54 <asselin> hi apoorvad
17:32:40 <apoorvad> Usually I miss the first half due to conflict. I would be able to chip in to testing as mmedvede mentioned
17:33:29 <mmedvede> asselin: thank you for the blueprints
17:34:01 <asselin> apoorvad, ok we're going to create a feature branch so facilitate testability.
17:34:10 <asselin> any preferences on branch name?
17:34:23 <apoorvad> asselin: cool. ciwatch-dev ?
17:35:51 <mmedvede> asselin: would it be temp branch just to implement initial tests?
17:36:17 <asselin> mmedvede, I can see it two ways
17:36:49 <asselin> we create the branch, and commit to it as usual. When we're happy, we merge it to master, or squash it, and proposed that as a patch to master.
17:37:21 <asselin> if we squash it, it would be temporary. If we don't....actually not sure if git keeps the branch name
17:38:06 <asselin> I guess it doesn't...so the name wouldn't stick. we can delete the branch after we merge it.
17:39:11 <mmedvede> asselin: I thought merges were a bit hard with gerrit
17:39:35 <asselin> I would prefer to have 'test' in the branch name, or something that hints at testability
17:39:40 <mmedvede> e.g. you can not review a merge commit
17:39:44 <mmedvede> unless it changed
17:39:57 <clarkb> mmedvede: can you expand on that? you definitely can review emrge commits
17:39:58 <mmedvede> asselin: +1 for more specific branch name
17:40:30 <asselin> clarkb, can we simply merge a feature branch to master in gerrit?
17:40:46 <clarkb> asselin: ya, you push it like any other change, it gets reviewed, and merges just like anything else
17:41:17 <asselin> clarkb, so you push the n commits of it? or something else?
17:41:20 <clarkb> there are two major considerations, the first is we don't allow pushing merge commits by default but can delegate that to individuals and the diff shows in the change for a merge commits shows you the conflicts iirc
17:41:35 <clarkb> asselin: well you push the n commits of it to branch testfoo, they merge there
17:41:45 <clarkb> asselin: then you merge testfoo into master and push that commit as a change against master
17:41:56 <mmedvede> clarkb: That is great I am wrong. Not sure where I got it from
17:42:22 <asselin> clarkb, and we need special permissions to do that?
17:42:33 <clarkb> asselin: you need special permissions to push merge commits yes
17:42:48 <clarkb> but we give those out when projects use feature branches
17:43:04 <mmedvede> clarkb: but do you need special permissions to submit merge commits for review?
17:43:14 <clarkb> mmedvede: yes
17:43:26 <clarkb> but thats not a big deal, we give them out regularly
17:44:01 <clarkb> that permission is not on by default because it helps avoid problems with people pushing up trees that have had implicit merges locally from eg git pull
17:45:27 <asselin> clarkb, and to create a feature branch, we just ask for on in -infra?
17:45:29 <asselin> one*
17:45:32 <clarkb> yes
17:46:49 <asselin> clarkb, thanks for the clarification
17:49:13 <asselin> ok, so back to branch names
17:49:15 <mmedvede> asselin: I suggests to name the branch "tests"
17:50:01 <asselin> or unit-tests
17:50:02 <mmedvede> and then merge/delete it once initial testing is implemented
17:50:09 <mmedvede> asselin: or unit-tests
17:50:10 <asselin> +1
17:50:17 <apoorvad> +1
17:50:28 <clarkb> the name should have a feature/ prefix to follow convention
17:50:39 <asselin> feature/unit-tests
17:50:48 <clarkb> probably less important here, but nice to follow anyways
17:52:26 <asselin> #action asselin to request -infra for feature/unit-tests branch created after merging any outstanding reviews we'd like to start with
17:53:27 <asselin> anything else on this topic?
17:54:03 <asselin> #topic Common-CI Solution
17:54:38 <asselin> #link Cannot run jobs on jenkins master: https://review.openstack.org/#/c/252768/
17:54:50 <asselin> #link Zuul cannot connect to openstack gerrit: https://review.openstack.org/#/c/266041/
17:55:51 <asselin> just needs some reviews ^^
17:56:10 <asselin> #topic open discussion
17:56:22 <ja3> nothing on my end
17:56:26 <asselin> anything else to discuss?
17:56:30 <mmedvede> asselin: was the zuul patch tested? (266041)
17:57:19 <mmedvede> I guess you are not the author :)
17:57:24 <asselin> I did not test it, but patrick-east did. I can test it quickly today
17:58:49 <asselin> ok, if nothing else we can end the meeting.
17:58:59 <mmedvede> thank you, asselin
17:59:10 <asselin> #endmeeting